Delete Delete Delete

[The_Almighty_Bacon_Lord]

Delete Delete Delete

[Rerouter]

Its generally that a bench / lab power supply is made for low noise. and to output a reliable voltage,

This can be done with a buck converter, but it comes down to how well the circuit is made to regulate over the entire desired output range
things like calculating the inductor values to output 2-30V at no load to 3A is hard, what iss good for one instance isnt that great for another

As for just providing crude power to devices that dont care that its noise, or a 0.5V difference wont harm, they are great, i have a box of step-up and step-down modules off ebay for times just need something close enough, but i also have 2 main lab supplies for when it does matter

In most cases when your just cobbling together a device, a clean and reliable supply removes one of the assumptions and makes it a bit easier to find out what went wrong elsewhere.

Also depending on your price point, a number of bulk buck converter supplies will kill output when shorted, while most bench supplies will fall back to a constant current mode, which again if you need it, it helps,

[Brumby]

Large output capacitors will hold a lot of energy at the set voltage.  These capacitors would need to be drained if there was a need for a lower terminal voltage.

For example, say you have a supply with:
 - current limiting set to 20mA
 - voltage set to 9V
 - a substantial amount of output capacitance

The you connect a red LED across the output.  This LED will first have to drain a chunk of energy from the output capacitors before the supply gets into current limiting.  This will happen quickly and is going to kill your LED.

[Rerouter]

The other thing you will find is most lab supplies with a constant current mode will have a very small output capacitor, for my 30V / 10A one, it only uses 47uF, but its so well built that i only see 35mV of ripple at full load (it is a switch mode based bench supply)

This is so the output voltage can fall very quickly when the current load changes to something greater than the limit.

In reality if you are intending to run something like an LED, you would set the current limit, turn the voltage limit to 0, connect the device and wind back up, as that output capacitor will always mean there will be some kind of spike when coming from a higher voltage, or at the very least set the voltage close to what you expect.

And i will warn you one thing that will bite you in the backside building your own, like i did mine, stability with non resistive loads, e.g. how does your supply react to being connected to car battery, or a large drill motor, or something that pulls spikes of current at a frequency,

[Kleinstein]

Without an filter, a switched mode regulator has quite some ripple. To reduce ripple one needs quite a large output capacitor.

Stability with non resistive loads is already the main difficulty in a linear lab supply. If gets even mode difficult with a switched mode supply, as the output stage is slower than just a BJT. This is why transient response of a switched mode supplies is usually not that good, even if well adjusted.

With the output capacitor one has to be a little careful when comparing. Those circuits with a small output capacitor usually use a emitter follower or similar output stage. Often the control circuit here is quite slow switching from CV to CC mode and this adds a behavior that is similar to having a rather large extra cap there. So even if the physical capacitance is small, there can be still quite a large current spike when going from CV to CC mode. So it is not only the capacitance you can see that is important, it is the one that one could measure for the CV to CC transition or CC mode. Quite often if there is little visible, a lot is hidden as a simulated capacitance.

[Mechatrommer]

using 2 quadrant topology somekind of push pull transistors output, that if more difficult load is to be taken into consideration, such as inductor (motor) and higher voltage battery. OTOH using 1 quadrant topology if only simple load is considered such as resistive load and lower voltage battery charging, usually found in beginner's diy psu, as simple as emitter follower topology.

[Kleinstein]

It depends on the type of supply, not all of them are the same. However the commercial supplies have the same limitations. So most of them do have a large output capacitance in one form or the other. There are a few variations:

1) Regulation with a current setting output stage, especially those with a floating regulator:
Such bench supplies usually have quite some output capacitance, like something in the 200 - 2000 µF. A few may be with something in the 100µF range. The faster the output stage, the smaller the output cap can be.

2) Supplies with a emitter follower (or similar) output stage and ground referenced regulation:
This type of regulator can get away with a low or even no capacitance at the output, but the circuit adds a virtual one. So there might be only a small capacitor, but if you check the CV-CC transition there still is quite some charge pulse, maybe like a capacitor in the 10 µF - 10 mF range. Also impedance CC mode is often not that high - more like a sizable capacitor, though often not as bad as the onset of CC mode.

3) There is also the Version with an low impedance output stage (e.g. emitter follower) for the voltage regulation and a floating regulator for the current limit. This could work with a low output capacitance and still have fast current limiting. Some plans based on the LM723 work this way, but with rather inaccurate current regulation. I don't know a new  commercial lab-supply that works this way. A few very old ones (MC1466 based) come close. Digital control and display makes the floating current regulator even less attractive.

There are also a few supplies that have small output cap, but are unstable with extreme loads because of this.
Having a 2 quadrant output stage can really help to make the output stage behave well at low currents as well and thus allow a smaller cap. However there are still limitations due to parasitic effects, especially parasitic inductance.

One just has to accept that lab supplies usually have that large capacitance. Once you know it is there,  it is usually not that bad.

[zapta]

Why would a buck converter make a bad power supply? I've heard a bit of information around the web that they usually tend to be really crappy, and that it shouldn't be done.
The reason I ask this is because I was planning on making a 30V 3A power supply with an XL4016 (datasheet: http://www.xlsemi.com/datasheet/xl4016%20datasheet.pdf)  IC, and have an input voltage of 32V, and an output voltage of 1.25V (idk how to make it 0V...?) to 30V, with an output current of close to 0A-3A.
Anyway, why do no "decent" bench power supplies (stuff around $50-200 on EBay) use buck converter IC's, and instead use mosphets or whatever else?

It's depend on the kind of things you do. I am using a switch mode bench power supply (Tenma 72-8350A) and it works well for me (hobby use, digital circuits).  Not every application require super low noise power supply. Often I also use a 5V phone charger to power my circuits and it works just fine.

[zapta]

I found this on Ebay: http://www.ebay.ca/itm/0-30V-2mA-3A-Adjustable-DC-Regulated-Power-Supply-DIY-Kits-LED-Display-Variable-/232058112709?hash=item3607bdcac5:g:RfoAAOSwHoFXvQRV

Would I be able to remove the bridge rectifier and a few other components and make it work on DC? Also, how would it compare to the original buck converter which I was planning on making with the XL4016?

I don't see why you can remove or bypass the bridge and run on DC (possibly even passing the DC  through the bridge even though the load will be on two diodes only).

There was a thread here a week or two ago about that board. Somebody was interested in getting just the PCB and this developed to a wider discussion.

You can also search eBay for similar boards that use switch mode converter.  Will simplify the power dissipation.

[Kleinstein]

The circuit from the Ebay links needs AC, as it uses a kind of charge pump to generate a low power negative supply.

Another thing to keep in mind is that often OPs are supplied that can not stand high voltage. So the safe upper limit for the input voltage is more like 18 V and thus maybe 22 V maximum at the output. To get the promised 30 V one would have to change the OPs to a better (more expensive) type that can stand higher supply voltage.

Also 3 A is rather optimistic for the single transistor. Though it can work with a 18 V or lower transformer. It still need really good cooling.